Atomistic Model of Realistic Crystalline Mesoporous Organosilica

Subscriber access provided by - Access paid by the | UCSB Libraries

C: Surfaces, Interfaces, Porous Materials, and Catalysis

Atomistic Model of Realistic Crystalline Mesoporous Organosilica Materials Including Nanochannels Antonio De Nicola, Andrea Correa, Angiolina Comotti, Piero Sozzani, and Giuseppe Milano J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b04645 • Publication Date (Web): 12 Jul 2018 Downloaded from http://pubs.acs.org on July 13, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Atomistic Model of Realistic Crystalline Mesoporous Organosilica Materials Including Nanochannels Antonio De Nicola,∗,† Andrea Correa,‡ Angiolina Comotti,¶ Piero Sozzani,¶ and Giuseppe Milano∗,† †Department of Organic Materials Science, Yamagata University, 4-3-16 Jonan Yonezawa, Yamagata-ken 992-8510, Japan. ‡Department of Chemical Sciences, Federico II University of Naples Via Cintia, Complesso Monte S. Angelo, 80126 Napoli, Italy. ¶Department of Materials Science and INSTM, University of Milano Bicocca, Via R.Cozzi 55, I-20125 Milan, Italy. E-mail: [email protected]; [email protected]

Abstract The new class of periodic mesoporous ogranosilicas materials (PMOs), due to the peculiar features, attracts growing interest from several research areas. We present an atomistic model of p-phenylenesilica crystalline mesoporous structure with a hexagonal framework, explicitly including channels on nanoscale. OPLS-AA force field optimization, to get suitable PMO structure compared with experimental one, is described. In particular, DFT calculation have been performed to calculate torsional energy barrier of phenyl rings connected to the silicon atoms belonging inorganic layers. Finally, the gas insertion and interaction of them with nanochannel walls are investigated.

1

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Introduction Periodic mesoporous organisilicas (PMOs) are a new class of materials 1 of particular interest due to the potential applications in several areas such as: gas storage, 2,3 confined polymerization, 4 molecular rotors, 3–6 selective adsorption, 3,7 catalysis, 8 molecular rotors, 9 adsorbent for pollutes from air and water, 10,11 and even water desalination. 12 The high technological impact of PMOs with large specific surface area makes the development of this materials a currently expanding research area. Since 1999 when Inagaki, 13 Asefa 14 and Melde 15 reported firsts successful synthesis of bridge-bonded species (RO)3 Si−R9 −Si(OR)3 , in which Si−C bonds are integrated into walls of a mesoporous silica matrix, the synthesis of hybrid organic-inorganic materials has become well established. Today, PMOs materials are commonly produced by hydrolytic and polycondensation reactions of precursors of type (RO)3 Si−R−Si(OR)3 known as sol-gel process. 16–20 By choosing specific R group, homogeneously within inorganic framework, a variety of features, such as surface hydrophobicity or accessible free volume, can be controlled. 2,21–23 Due to the peculiar structure, PMOs materials have unique features: chemically reactive sites homogeneously distributed, high of organic content; mesoporous channels; high surface area, high accessible free volume, fine tuning of the physical properties by chemical functionalization. In particular, PMOs with crystalline-like order in their walls 13–15,24 have organosiloxane units organized into ordered layers with molecular axes parallel to the mesochannel axis. Mesoporous organosilica crystalline structure exhibits a hexagonal framework with nanochannels having both, mesoscale periodicity and molecular order, as reported by Inagaki. 24 The combination of ordered organic/inorganic layers at wall nanochannel interface makes interactions of PMOs with guest species one of the key feature of adsorption/desorption properties from both, fundamental and applicative points of view. From the theoretical point of view and to the best of our knowledge, only a DFT study of PMOs investigating gas adsorption energy have been reported, 25 while, silica (SiO2 ) structures, crystalline and amorphous, 26–32 and organosilicate 26,33 have been extensively studied. 2


Page 2 of 29

Page 3 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In particular, DFT studies have been reported to investigate water adsorption on α-quartz surfaces, 31,32 while, combining Molecular Dynamics (MD) and the reactive bond-order based force field ReaxFF, 34 the stability of hydroxylated silica surface and crack propagation were investigated. 28–30 By using the same ReaxFF, Du et.al recently reported a study about nanoporouse organosilicate glasses 33 in which a nanoporouse silica systems were functionalized with methyl groups and short hydrocarbon chains. Our aim is to model and validate, at atomistic level and by using MD technique, a realistic structure of crystalline mesoporous organosilica. The paper is organized as following. In the next section the atomistic model and the force field optimization of crystalline precursor of PMOs, to get a correct description of the main interactions we will have in the hybrid mesoporous material, is reported. The procedure to get p-phenylenesilica crystalline mesoporous structure and force-field validation is reported in the Result and Discussion section. A first application investigating gas adsorption properties of PMO material is also given in the same section. The summary of the main results is given in the Conclusion section.

Methods and Model Due to the chemical similarity to the PMOs, OPLS-AA force field 35 have been tested to reproduce the PMOs crystalline precursor structure reported by Corriu. 36 On the basis of experimental X-Ray data, 36 the Cartesian coordinates of PMOs precursor crystalline structure have been generated by using Mercury CSD 2.0 Code. 37 In particular, a single unit cell of 1,4-Bis(trihidroxysilil) Benzene (BTHB) has been replicated five times in each x,y,z, direction, obtaining a periodic structure composed by 14400 atoms. The chemical structure of (BTHB) together with crystalline unit cell representation are reported in Figure 1. The consistency of the BTHB crystalline structure has been checked by comparison the calculated X-Ray pattern with those one experimentally reported. 36 We did not observe relevant differences between both X-Ray patterns, confirming that the obtained crystalline

3



Page 4 of 29

Figure 1: Chemical structure of BTHB on left. Crystalline unit cell, built from fractional atomic coordinates 36 in the right panel. Hydrogen atoms have been manually added by considering the bond distances and bending angle of the OPLS-AA force-field. 35 structure of BTHB is consistent with the experiments. OPLS-AA atomistic force-field 35 has been used for the majority of bonded and non bonded interactions, excluding those ones involving silicon atoms. In particular, bond and angle potential are calculated as: X

Ebond =

KB (r − req )2

(1)

Kθ (θ − θeq )2

(2)

bonds

X

Eangle =

angles

Where KB is the bond constant of the harmonic potential, req. is the equilibrium distance between two bonded atoms, Kθ is the angle bending constant and θ0 is the equilibrium angle between three consecutive bonded atoms. The torsional intra molecular energy is calculated as the sum of first three terms of Fourier series as:

E(φ) =

V2 V3 V1 [1 + cos(φ + f 1)] + [1 + cos(2φ + f 2)] + [1 + cos(3φ + f 2)] 2 2 2 4


(3)

Page 5 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Where φ is the dihedral angle between for consecutive bonded atoms and V1 ,V2 ,V3 are the Fourier coefficients series, instead f1, f2, f3 are the angle phases. The sum of Coulomb and Lennard-Jones (LJ) terms are used to represents the non-bonded interactions. The functional form of interaction energy for two molecules A and B is reported in the eq. 4:

EAB =

A X B X qi qj e 2 i

j

rij

+ 4ij

σij12 σij6 − 6 12 rij rij

fij

(4)

where fij is equal to 0.5 (for i, j = 1, 4) and is equal to 1 for i, j grater than 4. All bonded and non-bonded parameters used in this work are listed in the Table S1 of Supporting Information. The parameters involving silicon atoms, such as bond and angle parameters between Si−O, O−Si−O, Si−O−H groups, have been taken from the work of Barbier. 27 The initial system has been simulated for 1 ns in the N V T ensemble at 293 K and 1 atm by using GROMACS 4.5.4. 38 A time step of 2 f s has been used for all simulations. The temperature and pressure have been controlled by using Berendsen thermostat 39 with a τt of 10 f s, that ensure a reproduction of dynamic and structural properties at equilibrium. 40–42 Periodic boundary conditions have been applied to all x,y,z directions. After an initial energy minimization (500 steps) by using the steep-descent algorithm, all bonds have been constrained by using the LINCS algorithm. 43 The non-bonded interactions were truncated at 1.0 nm. After a short N V T simulation, a production run of 100 ns (N P T ensemble with τp = 30 f s) has been performed starting from the last N V T configuration. The composition of simulated systems is reported in Table 1.

Results and discussion By using the proposed model, we observed a high underestimation of the mass density, about + 28 %, together with a large deformation of the crystalline BTHB structure (Figure 2). 5



Table 1: Composition of Systems. System CO2 H2 O BTHB PMO Tot. Nr. of Atoms Temp. (K) time (ns) 1 0 0 250 0 14400 293.15 100.0 2 0 0 0 1 21600 293.15 100.0 3 760 0 0 1 23880 195.15 100.0 4 1168 0 0 1 25104 195.15 100.0 5 3976 0 0 1 33528 195.15 100.0 6 1000 1000 0 1 27600 195.15 100.0

Figure 2: Comparison between mass density calculated by using OPLS-AA (green open circles) and the tuned force-field (blue open cirlce). The black line represents the experimental 36 value of 1464 Kg/m3 .

6


Page 6 of 29

Page 7 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Considering that the OPLS-AA force-field 35 is able to reproduce the bulk mass density of organic liquid molecules within an error of about 2-4 %, and, considering that Barbier 27 reported a correct reproduction of the mass density and crystalline structure of silica αquartz, the main problem arises from the treatment of interactions between silicon atoms with phenyl carbon atoms. Due to the lack in literature, to the best of our knowledge, of parameters describing interactions between silicon and carbon atoms in a chemical equivalent BTHB structure, we decided to perform ab initio calculations to get suitable parameters of such interactions. In particular, on a single molecule of BTHB, DFT calculations have been performed to get bond lengths, bending and torsion angles. We found that the description of torsional energy profile between O−Si−C−C was critical in the correct reproduction of mass density and stability of structure of the BHTB crystal. From the energy profile of such torsional angle we obtained parameters, by mathematical fitting, to correctly describe both: energy barrier and shape of the torsional angle. As can be seen from Figure 3, the torsional barrier calculated from ab initio results to be very low, less than 1.0 Kcal/mol. The resulting energy profile has been fitted to fine tune original OPLS-AA parameters. We used the same functional of torsional energy (3) as fitting function. By using tuned parameters for O−Si−C−C torsional angle, a production run of 100 ns have been performed in the NPT ensemble, preceded by energy minimization and short NVT simulation, as described in the previous section. A comparison of mass density calculated from the tuned force-field with those ones obtained from non optimized OPLS-AA 35 and experimental data 27 is reported in Figure 2. We found that the tuned force-field reproduces the mass density within an error of 1.6 % with respect to experimental one. 27 Even if the tuned force-field well reproduces the mass density, the stability of the crystal structure is not still retained. In particular, distortions of both: cell length axes (a, b, c) and axes angles were observed. We have found that the treatment of the electrostatic interactions influences the stability of the crystalline cell parameters (a, b, c, α, β, γ). The stability of crystal structure has been tested over the long range electrostatic treat-

7



Figure 3: Comparison between the potential profile for O−Si−C−C torsional angle (right side) obtained by: DFT calculations (blue line), fitted function (red line), and, from MD simulation (black cross). DFT calculations have been performed using Gaussian09 package 44 at PBEH1PBE 45 level of theory; the electronic configuration of all atoms has been described by the 6-13G(d,p) basis set. 46–48

8


Page 8 of 29

Page 9 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ment. In particular, we tested the tuned force-field with a cut-off at 0.9 nm with charge groups, and the non optimized OPLS-AA with a default cut-off of 1.0 nm. As known from literature, 49,50 charge groups were introduced to reduce the artifact due to the cut-off truncation of electrostatic interactions. In particular, by using plain cut-off scheme for Coulomb interactions, relevant jumps in the forces and electrostatic potential arise when partially charged atoms move out and in the cut-off. To overcome this artifact and reduce jumps in Coulombic potential, chemical moieties having total zero charge (charge group) are considered to be in or out the neighbor list. The charge group scheme adopted for the proposed model is reported in the Supporting Information. Time behavior of cell parameters (a,b,c) obtained for different electrostatic treatment, together with snapshots of final configurations for OPLS-AA and tuned force-field, are reported in Figure 4. As can be seen from the figure, optimized force field and plain cut-off (0.9 nm with charge groups) for the electrostatic interactions, guarantees a satisfactory reproduction of cell parameters. In particular, the cell parameters are reproduced within an error of about 1.8-2.2% with respect to experimental ones. In addition, considering that we obtained similar results by using the Particle Mesh Ewald 49 , electrostatic treatment with plain cut-off is less expensive from the computational point of view with respect to PME (N log(N )). This aspect become relevant for applications, such as studies involving the dynamic of molecular rotors (on µs time scale), that are a peculiar future of PMO materials. it is worth noting that non optimized OPLS-AA overestimate the cell lengths of about 14-15%, and differently from the case of optimized force field, the crystalline order is lost, as can be seen from the snapshots of Figure 4B. In order to quantify how much the order is retained in the structure obtained from the tuned force-field, X-Ray pattern has been calculated and compared with the experimental one. 27 In Figure 5 the comparison between calculated and experimental X-Ray pattern is reported. The calculated X-Ray pattern clearly shows that the structure obtained from MD simulations has a comparable periodicity as the experimental one. 27 Moreover, the cell parameters

9



Figure 4: (A) Time behavior of cell parameters a,b,c calculated for non optimized OPLS-AA and tuned force-field in which electrostatic potential are treated with different settings. The segmented black lines represent experimental values. 27 All simulations have been performed in NPT ensemble in which the Berendsen barostat 39 have been used. Anisotropic pressure coupling have been applied for all simulations. (B) Snapshots for the ac plane of BTHB crystal for: initial crystal structure, non optimized OPLS-AA and tuned force-field.

10


Page 10 of 29

Page 11 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5: Comparison of X-Ray pattern between calculated (from NPT simulation with tuned force-filed) and experimental one 27 (red line). Numerical Hartree-Fock wave functions 51 have been used for the X-Ray pattern calculation. X-Ray calculation has been performed by averaging 20 configurations over the last 5 nanoseconds of NPT simulation.

11



have been reproduced within an error of about 1.8-2% and the monoclinic space group P 21 /c is correctly retained belong NPT MD simulations. In addition, the tuned force-field correctly reproduce the mass density within an error of about 1.6% with respect to experimental value. 27 The list of optimized parameters is reported in the Supporting Information. Model of Mesoporous Organosilica. The first example of complete crystal mesoporous organosilica has been reported by Inagaki et al., 1,13,18,19,24 which described the synthesis procedure to obtain the mesoporous p-phenylenesilica structure having a hexagonal array with a lattice constant of 5.25 nm and periodic channels of diameter of about 3.8 nm. In particular, the pore walls have been found to have a periodicity with a spacing of 0.76 nm along channel direction in which alternate organic/inorganic layers are connected by Si−C bonds. 24 A schematic model of the mesoporous p-phenylenesilica is reported in Figure 6. In the following is described the procedure to build the periodic mesoporous crystalline structure with hexagonal framework including explicitly nanochannels. In particular, we used as minimal brick element to build the whole structure, a single interconnected wall, as reported in Figure 7A. 12 replicas of the single wall (minimal brick element) have been oriented in a way to obtain a mesoporous repeating unit. Then, each wall has been properly connected, with others, by replacing pendant hydroxyl groups OH- with Si−O−Si covalent bonds. The final repeating unit, containing a whole nanochannel, has reported in the Figure 7B (continuous black line). Connecting 4 repeating units, as depicted in the scheme of Figure 7B, the fully connected periodic crystalline structure is obtained. A snapshot of the structure is reported in the panel C of Figure 7. The final mesoporouse structure, obtained by following the described procedure, has been used as starting point for MD simulations. In particular, a minimization of structure (250 steps), using steep-descent algorithm, have been performed. On the minimized structure a short NVT simulation of 5 ns has been performed followed by a production NPT run of 100 ns. For all simulations we used the tuned force-field validated in the previous section. In order to validate the structural properties of the mesoporouse crystalline structure,

12


Page 12 of 29

Page 13 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6: (A) Schematic model of Comparison of p-phenylenesilica. The wall periodicity of 0.76 nm and the channel diameter of 3.8 nm are indicated on the sketch. (B) Representation of of lattice model building process from the BTHB crystal structure. A slab of precursor PMOs crystal containing 22 phenyl rings, grouped in the black box, have been selected to be connected (by OH – removing) to form p-phenylenesilica lattice. Top and lateral views of the lattice are reported on the right of B panel. In yellow are represented silicon atoms, while, in red and green are reported oxygen and carbon atoms. Hydrogen atoms have been omitted for clarity. we compared with experiments 24 the distance distributions giving the main periodicity of the crystalline structure. In particular, we considered three main distances (D1, D2, D3) involving silicon atoms. D1 and D2, calculated between silicon atoms belonging same layer (Figure 8A), are responsible for the wall thickness, while, D3 is the distance between silicon atoms belonging two consecutive layers. Moreover, the D3 distance give us the periodicity of inorganic layers belong the mesoporous channels of the structure. The definitions and distributions of D1, D2 and D3 distances are reported in Figure 8. As can be seen from the figure, we have found that D1 and D3 distance distributions are peaked between 4.95.0 ˚ A, with a difference of about ≈ 0.5 ˚ Awith respect to experimental values. 24 Instead, from the distance distribution D3 it can be seen that a small range between 7.5-7.7 ˚ Aresults

13



Figure 7: (A) View along channel direction of both: single wall and assembled hexagonal crystalline unit. (B) Scheme illustrating the repeating unit (black continuous line) and periodic images along y,z plane (segmented black line). The top view of the repeating unit structure is also reported. (C) Top view of periodic mesoporous structure in which nanochannels are reported. In all reported structures the hydrogen atoms have been omitted for clarity. to be the most probable distance within a deviation of about ≈ 0.1-0.3 ˚ Awith respect to experimental 24 (7.6 ˚ A). Gas Inclusion. Due to the mesoporosity and high surface area 2,3,7 (up to 2730 m2 /g), gas adsorption capability (such as CO2 , CH4 , benzene, N2 and NH3 ) of PMOs has been recently measured. 3,7 In particular, IR 52–54 and solid-state NMR experiments 3 proved that p-phenylenesilica has high capability to adsorb CO2 molecules (up to 88% by weight at 195 K). In addition, DFT calculations have been performed to study interactions of CO, CH4 and CO2 molecules in a periodic wall of p-phenylenesilica. 25 In the following we report the simulation results of gas included in the pores, interacting with organic/inorganic layers, for four different systems. In particular, we simulated three systems in which CO2 ranges from 4.0 to 22.5% by weight, and, an additional system containing 5.6% of CO2 and 2.4% of H2 O, 14


Page 14 of 29

Page 15 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8: (A) Snapshots illustrating experimental distances 24 D1 and D2 between silicon atoms belonging the same layer of mesoporous structure. The distance distributions calculated from the NPT simulation are reported in the same panel. (B) View of mesoporous pore along channel parallel direction. The experimental distance 24 D3 between silicon atoms belonging two consecutive layers is reported. On the right side panel, calculated distance distribution D3 and the X-Ray pattern relative to the specific periodicity are reported. 15



respectively (systems 3-6 of Table 1). Gas molecules have been randomly inserted in the PMO nanochannel free volume by using packmol code. 55 Similar simulation protocol described in the previous section has been applied to all systems containing gases molecules. For CO2 molecule type OPLS-AA 35 parameters have been used, while, SPCE model 56 has been considered for H2 O molecule type. Moreover, the OPLS-AA force field has been extensively validated in the reproduction of structural and dynamic properties of hydrogen bonds. 57–62 In order to understand the effect of the CO2 concentration on the interaction with PMO walls, radial distribution functions (RDF) for different atom pairs have been calculated. A comparison of RDF is reported in Figure 9.

Figure 9: RDF calculated for the pairs: O(CO2) −Si(PMO) (black line), O(CO2) −OH(PMO) (red segmented line) and O(CO2) −C(CO2) (blue line). (A) System 3, (B) System 5 in Table 1. The RDF analysis has been performed by averaging configurations belonging the last 5 ns of each simulation. In particular, we compared RDFs for systems at 4.0 and 22.5%w/w of content of CO2 . In the first case (Figure 9A) it is clear as CO2 molecules interact more favorable with the pendant hydroxyl groups OH(PMO) exposed inside nanochannels of PMO compared to same CO2 molecules (blue line), in agreement with the result of Pacchioni 25 obtained from DFT calculations. In the second case, when the CO2 amount increases (22.5%(w/w) ), the self inter16


Page 16 of 29

Page 17 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


actions between CO2 −CO2 molecules became more probable with the respect to those ones CO2 −OH(PMO) , as reported in Figure 9 B. Moreover, the peak at ≈ 0.6 nm of CO2 −CO2 (blue curve) confirms the presence of CO2 second shell molecules only at high carbon dioxide content. Using geometrical criteria, based on position and width of peaks in RDF calculated for CO2 −CO2 pair, molecules belonging first and second shell have been grouped. A representation of the two shell molecule types is sketched in Figure 10.

Figure 10: Snapshot of PMO configuration at 22.5%w/w of CO2 . The left side of the snapshot shows PMO structure in yellow and two different CO2 molecule types, in green belonging the first shell and in blue belonging to second shell type. On the right side of the figure a standard visualization based on different atom types is reported. Hydrogen atoms have been not included for clarity. The presence of a CO2 second shell molecule has effect also on the self diffusion coef17



Page 18 of 29

ficient D of carbon dioxide. D were calculated from the linear fitting of the mean square displacement of the Einstein relation 49 : 1 h|ri (t) − ri (0)|2 i t→∞ 6t

D = lim

(5)

The results including global and components along x, y, z directions of the diffusion coefficients are reported in Table 2. As general result, as expected, the diffusion is significantly reduced in confined y, z directions (see panel B of Figure 7). Instead, along non confined direction parallel to the nanochannel axis, Dx result to be ≈ 2 times larger. Moreover, the formation of a stable second shell (at high CO2 content, system 5 Table 2), reduces the mobility of carbon dioxide molecules, of a factor about 2, along confined and non confined directions. Time behavior of MSD are reported in the Supporting Information.

Table 2: Global and x, y, z components of D (10−5 cm2 s−1 ) calculated for CO2 molecules. System CO2 %(w/w) 3 4.0 4 6.6 5 22.5

D Dx Dy Dz 0.076 (0.008) 0.12 (0.01) 0.046 (0.007) 0.060 (0.001) 0.076 (0.004) 0.128 (0.001) 0.051 (0.001) 0.049 (0.006) 0.035 (0.001) 0.056 (0.005) 0.026 (0.001) 0.025 (0.001)

We also investigated difference in the gas/PMO wall interaction for CO2 and H2 O molecules. To this aim an additional system at 1:1 CO2 :H2 O (system 6 in Table 1) were simulated. The RDF comparison in Figure 11A shows a slightly stronger interaction of CO2 in favor of PMO compared to H2 O. Moreover, a stronger interaction between H2 O-CO2 , with the respect to the H2 O-hydroxyl groups belonging PMO walls is found (Figure 11B). This behavior were confirmed, also, comparing the hydrogen bond (H-Bond) fraction between H2 O and both: PMO hydroxyl groups and CO2 molecules, as reported in Figure 12A. In particular, we have found water molecules to form H-bonds more probable with CO2 molecules instead of pendant hydroxyl groups of PMO. Moreover, no relevant differences between donor-acceptor distance distributions and hydrogen bond lifetimes for the two different

18


Page 19 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 11: RDF calculated for the pairs: (A) O(CO2 ) −OH(PMO) (black line), O(H2 O) −OH(PMO) (red segmented line), and (B) H(H2 O) −O(PMO) (black line), H(H2 O) −O(CO2 ) (red segmented line). The RDF analysis has been performed by averaging configurations belonging the last 5 ns of the simulation. H-bonds (with CO2 and HO-PMO) have been found (Figure 12B, C). The presence of high number of H-bonds between CO2 and H20 influences also dynamic properties. In fact, do to formation of a H-bond network, the diffusion of CO2 molecules is found to be reduced of a factor ≈ 1.7 even the %(w/w) is low (5.6), as reported in Table 3.

Table 3: Global and x, y, z components of D (10−5 cm2 s−1 ) calculated for CO2 and H2 O molecules (system 6 of Table 1). Molecule D Dx Dy Dz CO2 0.043 (0.001) 0.642 (0.003) 0.027 (0.005) 0.03 (0.01) H2 O 0.003 (0.001) 0.004 (0.001) 0.002 (0.001) 0.002 (0.001)

Conclusions We provided a validation and optimization of OPLS-AA force field to model at all atom level periodic crystalline mesoporous p-phenylenesilica. In particular, the mesoporous hexagonal framework structure, including explicitly nanochannels, have been validated on the reproduc19



Figure 12: (A) Time behavior of H-Bond fraction. (B) Distribution of Donor-Acceptor distance, defined as the distance between AO(A-H-O). (C) Hydrogen bond lifetime. The H-Bonds have been calculated considering as geometrical criteria 63 a cut-off distance of 0.3 nm and angle between acceptor-donor groups A-H-O θ = 120◦ . Oxygen atoms of both: carbon dioxide and hydroxyl group of PMO are considered hydrogen bond acceptor A, while, H-O groups of water molecules are considered as donors. tion of experimental periodic distance distributions. 24 DFT calculations have been used to correctly parametrize torsional potential of phenyl rings belonging organic layers connected with inorganic silicon atoms. Furthermore, gas inclusion and gas/PMO wall interactions have been investigated for CO2 and H2 O molecules. No significant differences have been found for these two gases with only a slightly stronger interactions of CO2 /PMOs nanochannel with respect to water molecules. Instead, dynamic properties, such as diffusion coefficient D, has been found to be dependent from gas amount and composition. In the first case, at high CO2 content, a mobility reduction of carbon dioxide molecules is observed due to the formation of a second shell of CO2 molecules. Also in the second case the diffusion of carbon dioxide molecules is reduced thanks to the formation of a strong hydrogen bond network between CO2 and H2 O molecules.

Acknowledgement G.M and A.D.N. wish to thank the HPC team of Enea (http://www.enea.it) for using the ENEA-GRID and the HPC facilities CRESCO (http://www.cresco.enea.it) in Portici. 20


Page 20 of 29

Page 21 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A.C. would like to thank Cariplo Foundation (Balance project), PRIN 2016 and INSTM Consortium/Lombardy Region for financial support.

Supporting Information Available Supporting Information includes: optimized parameters for PMO structure, adopted charge and group charge scheme, additional plot of MSD for all simulated systems including gases in the nanochannel.

References (1) Mizoshita, N.; Tani, T.; Inagaki, S. Syntheses, properties and applications of periodic mesoporous organosilicas prepared from bridged organosilane precursors. Chem. Soc. Rev. 2011, 40, 789–800. (2) Xia, Y.; Mokaya, R. High surface area ethylene-bridged mesoporous and supermicroporous organosilica spheres. Microporous and Mesoporous Materials 2005, 86, 231 – 242. (3) Comotti, A.; Bracco, S.; Valsesia, P.; Ferretti, L.; Sozzani, P. 2D Multinuclear NMR, Hyperpolarized Xenon and Gas Storage in Organosilica Nanochannels with Crystalline Order in the Walls. Journal of the American Chemical Society 2007, 129, 8566–8576. (4) Angiolina, C.; Silvia, B.; Mario, B.; Jacopo, P.; Mauro, G.; Piero, S. Confined Polymerization in Highly Ordered Mesoporous Organosilicas. Chemistry - A European Journal 2015, 21, 18209–18217. (5) Angiolina, C.; Silvia, B.; Patrizia, V.; Mario, B.; Piero, S. Fast Molecular Rotor Dynamics Modulated by Guest Inclusion in a Highly Organized Nanoporous Organosilica. Angewandte Chemie International Edition 2010, 49, 1760–1764. 21



(6) Bracco, S.; Comotti, A.; Valsesia, P.; Chmelka, B. F.; Sozzani, P. Molecular rotors in hierarchically ordered mesoporous organosilica frameworks. Chemical Communications 2008, 4798–4800. (7) Wei, Y.; Li, X.; Zhang, R.; Liu, Y.; Wang, W.; Ling, Y.; El-Toni, A. M.; Zhao, D. Periodic Mesoporous Organosilica Nanocubes with Ultrahigh Surface Areas for Efficient CO2 Adsorption. Scientific Reports 2016, 6, 20769 EP –, Article. (8) Frank, H.; Maximilian, C.; J¨ urgen, M.; Michael, F. Silica-Based Mesoporous OrganicInorganic Hybrid Materials. Angewandte Chemie International Edition 2006, 45, 3216– 3251. (9) Bracco, S.; Beretta, M.; Cattaneo, A.; Comotti, A.; Falqui, A.; Zhao, K.; Rogers, C.; Sozzani, P. Dipolar Rotors Orderly Aligned in Mesoporous Fluorinated Organosilica Architectures. Angewandte Chemie International Edition 54, 4773–4777. (10) Wendong, W.; E., L. J.; A., O. G. Why PMO?: Towards Functionality and Utility of Periodic Mesoporous Organosilicas. Small 2010, 6, 2634–2642. (11) Chen, C.; Kim, J.; Ahn, W.-S. CO2 capture by amine-functionalized nanoporous materials: A review. Korean Journal of Chemical Engineering 2014, 31, 1919–1934. (12) Chua, Y. T.; Lin, C. X. C.; Kleitz, F.; Zhao, X. S.; Smart, S. Nanoporous organosilica membrane for water desalination. Chemical Communications 2013, 49, 4534–4536. (13) Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. Novel Mesoporous Materials with a Uniform Distribution of Organic Groups and Inorganic Oxide in Their Frameworks. Journal of the American Chemical Society 1999, 121, 9611–9614. (14) Asefa, T.; MacLachlan, M. J.; Coombs, N.; Ozin, G. A. Periodic mesoporous organosilicas with organic groups inside the channel walls. Nature 1999, 402, 867 EP –.

22


Page 22 of 29

Page 23 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(15) Melde, B. J.; Holland, B. T.; Blanford, C. F.; Stein, A. Mesoporous Sieves with Unified Hybrid Inorganic/Organic Frameworks. Chemistry of Materials 1999, 11, 3302–3308. (16) Hoffmann, F.; Cornelius, M.; Morell, J.; Fr¨oba, M. Silica-Based Mesoporous Organic– Inorganic Hybrid Materials. Angewandte Chemie International Edition 2006, 45, 3216– 3251. (17) Loy, D. A.; Shea, K. J. Bridged Polysilsesquioxanes. Highly Porous Hybrid OrganicInorganic Materials. Chemical Reviews 1995, 95, 1431–1442. (18) Fujita, S.; Inagaki, S. Self-Organization of Organosilica Solids with Molecular-Scale and Mesoscale Periodicities. Chemistry of Materials 2008, 20, 891–908. (19) Kapoor, M. P.; Yang, Q.; Inagaki, S. Self-Assembly of Biphenylene-Bridged Hybrid Mesoporous Solid with Molecular-Scale Periodicity in the Pore Walls. Journal of the American Chemical Society 2002, 124, 15176–15177. (20) Hunks, W. J.; Ozin, G. A. Challenges and advances in the chemistry of periodic mesoporous organosilicas (PMOs). Journal of Materials Chemistry 2005, 15, 3716–3724. (21) Jonas, C.; Xavier, C.; Chi, M. M. W.; Audrey, G.; Laurence, R.; Philippe, T.; Marie, M.; Jean-Olivier, D. Biodegradable Ethylene-Bis(Propyl)Disulfide-Based Periodic Mesoporous Organosilica Nanorods and Nanospheres for Efficient In-Vitro Drug Delivery. Advanced Materials 2014, 26, 6174–6180. (22) Fangqiong, T.; Linlin, L.; Dong, C. Mesoporous Silica Nanoparticles: Synthesis, Biocompatibility and Drug Delivery. Advanced Materials 2012, 24, 1504–1534. (23) Park, S. S.; Santha Moorthy, M.; Ha, C.-S. Periodic mesoporous organosilicas for advanced applications. Npg Asia Materials 2014, 6, e96. (24) Inagaki, S.; Guan, S.; Ohsuna, T.; Terasaki, O. An ordered mesoporous organosilica hybrid material with a crystal-like wall structure. Nature 2002, 416, 304. 23



(25) Martinez, U.; Pacchioni, G. Interaction of CO, CO2 and CH4 with mesoporous organosilica: Periodic DFT calculations with dispersion corrections. Microporous and mesoporous materials 2010, 129, 62–67. (26) Rimsza, J. M.; Yeon, J.; van Duin, A. C. T.; Du, J. Water Interactions with Nanoporous Silica: Comparison of ReaxFF and ab Initio based Molecular Dynamics Simulations. The Journal of Physical Chemistry C 2016, 120, 24803–24816. (27) Barbier, D.; Brown, D.; Grillet, A.-C.; Neyertz, S. Interface between End-Functionalized PEO Oligomers and a Silica Nanoparticle Studied by Molecular Dynamics Simulations. Macromolecules 2004, 37, 4695–4710. (28) Rimsza, J. M.; Jones, R. E.; Criscenti, L. J. Surface Structure and Stability of Partially Hydroxylated Silica Surfaces. Langmuir 2017, 33, 3882–3891. (29) M., R. J.; E., J. R.; J., C. L. Crack propagation in silica from reactive classical molecular dynamics simulations. Journal of the American Ceramic Society 101, 1488–1499. (30) Jincheng, D.; N., C. A. Molecular Dynamics Simulation of the Structure and Hydroxylation of Silica Glass Surfaces. Journal of the American Ceramic Society 2005, 88, 2532–2539. (31) Bandura, A. V.; Kubicki, J. D.; Sofo, J. O. Periodic Density Functional Theory Study of Water Adsorption on the a-Quartz (101) Surface. The Journal of Physical Chemistry C 2011, 115, 5756–5766. (32) Skelton, A. A.; Wesolowski, D. J.; Cummings, P. T. Investigating the Quartz (10¯10)/Water Interface using Classical and Ab Initio Molecular Dynamics. Langmuir 2011, 27, 8700–8709. (33) Rimsza, J. M.; Deng, L.; Du, J. Molecular dynamics simulations of nanoporous

24


Page 24 of 29

Page 25 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


organosilicate glasses using Reactive Force Field (ReaxFF). Journal of Non-Crystalline Solids 2016, 431, 103–111. (34) Yeon, J.; van Duin, A. C. T. ReaxFF Molecular Dynamics Simulations of Hydroxylation Kinetics for Amorphous and Nano-Silica Structure, and Its Relations with Atomic Strain Energy. The Journal of Physical Chemistry C 2016, 120, 305–317. (35) Jorgensen, W. L.; Maxwell, D. S.; Tirado Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. Journal Of The American Chemical Society 1996, 118, 11225–11236. (36) Cerveau, G.; Corriu, R. J. P.; Dabiens, B.; Le Bideau, J. Synthesis of Stable Organo(bissilanetriols): X-Ray Powder Structure of 1,4-Bis(trihydroxysilyl)benzene. Angewandte Chemie International Edition 2000, 39, 4533–4537. (37) Macrae, C. F.; Bruno, I. J.; Chisholm, J. A.; Edgington, P. R.; McCabe, P.; Pidcock, E.; Rodriguez-Monge, L.; Taylor, R.; van de Streek, J.; Wood, P. A. Mercury CSD 2.0 new features for the visualization and investigation of crystal structures. Journal of Applied Crystallography 2008, 41, 466–470. (38) Van der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A. E.; Berendsen, H. J. C. GROMACS: Fast, flexible, and free. Journal of Computational Chemistry 2005, 26, 1701–1718. (39) Berendsen, H. J. C.; Postma, J. P. M.; Gunsteren, W. F. v.; DiNola, A.; Haak, J. R. Molecular dynamics with coupling to an external bath. Journal of Chemical Physics 1984, 81, 3684–3690. (40) Basconi, J. E.; Shirts, M. R. Effects of Temperature Control Algorithms on Transport Properties and Kinetics in Molecular Dynamics Simulations. Journal of Chemical Theory and Computation 2013, 9, 2887–2899.

25



(41) Wong-ekkabut, J.; Karttunen, M. Assessment of Common Simulation Protocols for Simulations of Nanopores, Membrane Proteins, and Channels. Journal of Chemical Theory and Computation 2012, 8, 2905–2911. (42) Mudi, A.; Chakravarty, C. Effect of the Berendsen thermostat on the dynamical properties of water. Molecular Physics 2004, 102, 681–685. (43) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. LINCS: A linear constraint solver for molecular simulations. Journal of Computational Chemistry 1997, 18, 1463–1472. (44) Frisch, M. J. et al. Gaussian 09 Revision C.01. Gaussian, Inc.: Wallingford, CT, 2009. (45) Ernzerhof, M.; Perdew, J. P. Generalized gradient approximation to the angle- and system-averaged exchange hole. The Journal of Chemical Physics 1998, 109, 3313– 3320. (46) Ditchfield, R.; Hehre, W. J.; Pople, J. A. SelfConsistent MolecularOrbital Methods. IX. An Extended GaussianType Basis for MolecularOrbital Studies of Organic Molecules. The Journal of Chemical Physics 1971, 54, 724–728. (47) Petersson, G. A.; Bennett, A.; Tensfeldt, T. G.; AlLaham, M. A.; Shirley, W. A.; Mantzaris, J. A complete basis set model chemistry. I. The total energies of closedshell atoms and hydrides of the firstrow elements. The Journal of Chemical Physics 1988, 89, 2193–2218. (48) Petersson, G. A.; AlLaham, M. A. A complete basis set model chemistry. II. Openshell systems and the total energies of the firstrow atoms. The Journal of Chemical Physics 1991, 94, 6081–6090. (49) Allen, M. P.; Tildesley, D. J. Computer simulation of liquids; Oxford university press, 2017. 26


Page 26 of 29

Page 27 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(50) Frenkel, D.; Smit, B. Understanding molecular simulation: from algorithms to applications; Elsevier, 2001; Vol. 1. (51) Cromer, D. T.; Mann, J. B. X-ray scattering factors computed from numerical HartreeFock wave functions. Acta Crystallographica Section A 1968, 24, 321–324. (52) Hoffmann, F.; Gngerich, M.; Klar, P. J.; Frba, M. Vibrational Spectroscopy of Periodic Mesoporous Organosilicas (PMOs) and Their Precursors: A Closer Look. The Journal of Physical Chemistry C 2007, 111, 5648–5660. (53) Onida, B.; Borello, L.; Busco, C.; Ugliengo, P.; Goto, Y.; Inagaki, S.; Garrone, E. The Surface of Ordered Mesoporous BenzeneSilica Hybrid Material: An Infrared and ab Initio Molecular Modeling Study. The Journal of Physical Chemistry B 2005, 109, 11961–11966. (54) Camarota, B.; Ugliengo, P.; Garrone, E.; Arean, C. O.; Delgado, M. R.; Inagaki, S.; Onida, B. IR and Computational Characterization of CO Adsorption on a Model Surface, the Phenylene Periodic Mesoporous Organosilca with Crystalline Walls. The Journal of Physical Chemistry C 2008, 112, 19560–19567. (55) Martnez, L.; Andrade, R.; Birgin, E. G.; Martnez, J. M. PACKMOL: A package for building initial configurations for molecular dynamics simulations. Journal of Computational Chemistry 30, 2157–2164. (56) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. The missing term in effective pair potentials. The Journal of Physical Chemistry 1987, 91, 6269–6271. (57) Rizzo, R. C.; Jorgensen, W. L. OPLS All-Atom Model for Amines: Resolution of the Amine Hydration Problem. Journal of the American Chemical Society 1999, 121, 4827– 4836.

27



(58) Duffy, E. M.; Jorgensen, W. L. Prediction of Properties from Simulations: Free Energies of Solvation in Hexadecane, Octanol, and Water. Journal of the American Chemical Society 2000, 122, 2878–2888. (59) Lei, Y.; Li, H.; Pan, H.; Han, S. Structures and Hydrogen Bonding Analysis of N,NDimethylformamide and N,N-DimethylformamideWater Mixtures by Molecular Dynamics Simulations. The Journal of Physical Chemistry A 2003, 107, 1574–1583. (60) Hladlkov, J.; Fischer, H. E.; Jungwirth, P.; Mason, P. E. Hydration of Hydroxyl and Amino Groups Examined by Molecular Dynamics and Neutron Scattering. The Journal of Physical Chemistry B 2015, 119, 6357–6365. (61) De Nicola, A.; Correa, A.; Milano, G.; La Manna, P.; Musto, P.; Mensitieri, G.; Scherillo, G. Local Structure and Dynamics of Water Absorbed in Poly(ether imide): A Hydrogen Bonding Anatomy. The Journal of Physical Chemistry B 2017, 121, 3162– 3176. (62) De Nicola, A.; Avolio, R.; Della Monica, F.; Gentile, G.; Cocca, M.; Capacchione, C.; Errico, M. E.; Milano, G. Rational design of nanoparticle/monomer interfaces: a combined computational and experimental study of in situ polymerization of silica based nanocomposites. RSC Adv. 2015, 5, 71336–71340. (63) Van der Spoel, D.; van Maaren, P. J.; Larsson, P.; Tˆımneanu, N. Thermodynamics of Hydrogen Bonding in Hydrophilic and Hydrophobic Media. The Journal of Physical Chemistry B 2006, 110, 4393–4398.

28


Page 28 of 29

Page 29 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Graphical TOC Entry

29


Atomistic Model of Realistic Crystalline Mesoporous Organosilica

Recommend Documents